2056 XG IMDS

Last updated
Save as PDF

LexisNexis® Intelligent Match Decision Solution (IMDS)

R E P A P E T I H W

in LexisNexis® Bridger Insight® XG (XG)

L A C I N H C E T

BIS249-EN_US — 29 September 2023

CONFIDENTIAL

The recipient of this material (hereinafter "the Material") acknowledges that it contains confidential and proprietary data the disclosure to, or use of which by, third parties will be damaging to LexisNexis Risk Solutions and its affiliated companies (hereinafter "LexisNexis Risk Solutions"). Therefore, recipient agrees to hold the Material in strictest confidence, not to make use of it other than for the purpose for which it is being provided, to release it only to employees requiring such information, and not to release or disclose it to any other party. Upon request, recipient will return the Material together with all copies and modifications, if any, to LexisNexis Risk Solutions.

LexisNexis Risk Solutions shall not be liable for technical or editorial errors or omissions contained herein. The information in this publication is subject to change without notice and is provided "as-is" without a warranty of any kind. Nothing herein should be construed as constituting a warranty, as any applicable warranty is exclusively contained in your signed agreement with LexisNexis Risk Solutions.

All names, addresses, birth dates or other identifying information in the text, or on the sample reports and screens shown in this document, are of fictitious persons and entities and have been created for illustrative purposes only. Any similarity to the name of any real person, school, business, other entity, address, date of birth, or other identifying information is purely coincidental.

Due to the nature and origin of public record information, the public records and commercially available data sources used in reports may contain errors.

The LexisNexis Intelligent Match Decision Solution service is not provided by "consumer reporting agencies," as that term is defined in the Fair Credit Reporting Act (15 U.S.C. 1681, et seq.) (FCRA), and does not constitute "consumer reports," as that term is defined in the FCRA. Accordingly, the Intelligent Match Decision Solution service may not be used in whole or in part as a factor in determining eligibility for credit, insurance, employment or for any other eligibility purpose that would qualify it as a consumer report under the FCRA.

LexisNexis and the Knowledge Burst logo are registered trademarks of RELX Inc. Bridger Insight is a registered trademark of LexisNexis Risk Solutions Inc. WorldCompliance is a registered trademark of World Compliance, Inc. Other products and services may be trademarks or registered trademarks of their respective companies.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 2

CONFIDENTIAL

Con t ent s

Intelligent Match Decision Solution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Alert Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Implementation in Bridger Insight XG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Rule Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Intercept. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Rules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Coefficients. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

Reason Codes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13

False Positive Confidence Score. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Rule Examples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

BIS249-EN_US — 29 September 2023

IMDS in XG Page 3

CONFIDENTIAL

Intell ig ent Match Decision Solution

LexisNexis® Intelligent Match Decision Solution (IMDS) provides functionality to help refine your processing of automatic false positives in LexisNexis® Bridger Insight® XG (XG).

XG is designed to help organizations of all sizes and across various industries ensure that they are not engaging in business with prohibited entities. To accomplish this task, XG compares input entities to screening list entities. Some comparisons find true matches in which the input entity is the same entity as the list entity. However, some comparisons find false positives in which the input entity is similar to the list entity, but they are not actually the same entity.

In an ideal setting, every match would be a true match, because each entity would have an identifying data element that no other entity shares (for example, an SSN (Social Security number) in the United States). Unfortunately, false positives are common because some entity identifiers are not unique. For example, common names or street addresses that are occupied by multiple individuals can result in false positives.

Even a small percentage of false positives can quickly add up as the number of searched records increases. Your staff can spend many hours researching these matches, instead of spending time on strategic tasks and high-value activities. Rather than requiring staff to manually process (or adjudicate) every false positive, XG offers IMDS to help you minimize the impact from false positives in your list screening process.

IMDS lets your organization develop rules to automatically process list matches in a way that aligns with your organization's list screening program and risk profile. The rules leverage input, list, and match attributes from XG searches to test for evidence of a false positive. IMDS uses the results to determine a false positive confidence score. Your organization can specify alert processing actions for XG based on the false positive confidence score.

For example, maybe your organization considers matches to be false positives if they contain two or more data element conflicts, such as address, citizenship, country, or gender. IMDS rules can be written that test for each of those possible data element conflicts, and weight them so that any two or more true rules indicate full confidence of a false positive. Or, maybe your organization considers low name match scores and large DOB (date of birth) distances to be evidence of a false positive, and high name match scores and exactly matching DOB values to be evidence of a true match. IMDS rules can be written that cover a range of name score and DOB distance values, and progressively weight them depending on how strong the evidence of a false positive is for the particular rule. You can also leverage statistical modeling of previously-reviewed alerts to predictively determine rule weights.

You use IMDS to further customize your predefined search settings for automatic false positives. IMDS can help reduce the number of false positives and the number of matches within alerts, providing you with higher quality alerts.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 4

CONFIDENTIAL

By developing rules that are based on your organization's existing list screening program and risk profile, you can automate and standardize some of the manual, repetitive, and time-consuming tasks that divert resources away from high-value activities. IMDS helps you prioritize the most important alerts and focus on strategic tasks and key decisions.

Considerations

To help ensure the best outcomes for an IMDS implementation, your project team should understand your organization's list screening program and risk profile. Your project team also should be able to develop valid XML code, assess data quality, and validate the operational soundness of the rules.

Prior to the IMDS implementation, your project team can assess data quality (for example, the impact of data that is incomplete, data that contains duplicate records or truncated entries, and other data inconsistencies), evaluate historical alert decisions, and manage the impact of changes to your workflow and current configurations to manage false positive and false negative risks.

A well-planned, tuned, and validated implementation of IMDS can result in increased productivity and reduced operational costs. FCC (financial crimes compliance) Professional Services from LexisNexis® Risk Solutions can assist with the IMDS implementation and ongoing model monitoring, tuning, and validation processes. For more information, contact your sales representative.

Alert Management

In a predefined search, you can specify the rules to test a match for evidence of a false positive and how XG processes any resulting automatic false positives in alerts.

When only the input entity name matches a list entity name, XG offers standard rules for automatic false positives that you configure directly in a predefined search. The standard rules let you specify several possible conflicts between data elements for the input entity and the list entity. These conflicts include DOB within a specified tolerance in months, ID numbers, gender, addresses, countries, phone numbers, entity type, and citizenship. When any of the standard rules that you specify return as TRUE for a match, XG sets the match status to Automatic False Positive. However, if the addresses, DOB, ID numbers, or phone numbers match, then XG overrides the standard rule, and the match status is not set to Automatic False Positive.

The standard rules are designed with a limited scope that works well for many organizations. For example, if you specify the gender conflict rule and the entity type conflict rule for a predefined search, then either rule sets an Automatic False Positive status. However, if one of the override data elements matches, such as the addresses, then the Automatic False Positive status is not set, even if both the genders and the entity types conflict.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 5

CONFIDENTIAL

Rules for IMDS offer your organization a greater ability to combine input, list, and match attributes to help improve the quality of alerts. You can develop rules that align with your organization's existing list screening program and risk profile, because IMDS can directly use a wide variety of attributes, including and beyond data element conflicts. You can also specify a weight for each rule to change how strongly the rule affects the evidence for a false positive, which means that rules build on each other to provide a more comprehensive view of the evidence of a false positive that a match may contain.

You should not use the standard rules for automatic false positives in combination with rules for IMDS.

When you use rules for IMDS, you can also specify a confidence threshold that a match must meet to be an automatic false positive. IMDS calculates a false positive confidence score from the results of the rules and compares the false positive confidence score against the confidence threshold. The false positive confidence score can range from 0.0000 (no confidence that the match is a false positive) to 10.0000 (full confidence that the match is a false positive).

The false positive confidence score is not the same as the match confidence score (70-100) that XG provides for each match to indicate how closely the list entity matches the input entity.

Depending on how you configure the predefined search, XG can suppress matches with a false positive confidence score that meets the confidence threshold, or include the match in the alert and provide intelligent match decision information, including the false positive confidence score, reason codes, and the version of the rules. This information can help provide deeper insight into the false positive determination process and provide a record of evidence of a false positive in an audit trail.

A false positive confidence score of 10.0000 is not a guarantee that a match is a false positive. The score is a result of the rules that you develop from your organization's existing list screening program and risk profile, and may reduce the number of reported matches.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 6

CONFIDENTIAL

Imple ment ation in Bridger Insight XG

To implement IMDS in XG, the rules must be made available to client instances and configured in predefined searches. You develop rules in an XML file that is uploaded to the XG system.

XG Service clients provide the rule file to LexisNexis Risk Solutions to be set up in their client instance. XG Enterprise clients set up rules for client instances in Enterprise Manager.

Navigation: System > Intelligent Match Decision Rules > Add Rules

The product settings for each client instance indicate whether users can include IMDS in list screening searches. In the client instance, IMDS needs to be configured in the appropriate predefined searches.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 7

CONFIDENTIAL

In a predefined search, you configure IMDS as part of the rules for automatic false positives. You set the rules to use and the confidence threshold that a match must meet to be a reported as an automatic false positive.

Navigation: Administration > Predefined Searches > (View a predefined search) > (Expand the List Screening section) > (Click Automatic False Positive Rules)

Depending on the rules that are configured in the predefined search, when XG generates an alert, matches that meet the confidence threshold include an intelligent match decision. The intelligent match decision contains the following information:

• The false positive confidence score that IMDS calculated for the match • The reason codes that were returned for the match • The rules that were used to determine that the match was an automatic false positive

The intelligent match decision helps you understand which rules found evidence of a false positive and can provide a record of the evidence of a false positive in the audit trail.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 8

CONFIDENTIAL

The following image shows the intelligent match decision for a match that contained a citizenship conflict and a gender conflict.

Automatic False Positive Match With Intelligent Match Decision

An alert can contain multiple list matches. You can configure how XG reports these matches.

When all the list matches in the alert are automatic false positives, XG generates an alert by default with an open alert state and an Automatic False Positive alert status. You can configure the product to not generate an alert or to generate an alert with a closed alert state. If intelligent match decisions are configured, then XG also sets the match indicator to the false positive option.

When only some list matches in the alert are automatic false positives, XG generates an alert by default that contains only the list matches that are not automatic false positives. You can configure the product to also include the list matches that are automatic false positives.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 9

CONFIDENTIAL

Rule F i les

An XML rule file contains the rules that you develop for IMDS and upload to XG. A rule file contains an intercept and at least one rule. When XG returns a match, IMDS uses every rule in the rule file to test the match for evidence of a false positive, and adds each coefficient for a true rule to the intercept that the rule file contains.

You can implement multiple rule files in XG. For example, you may choose to upload one rule file with rules that are specific to an individual entity search, and upload a second rule file with rules that are specific to a business entity search.

Intercept

An intercept is the baseline value that you choose for a rule file to include in the calculation of the false positive confidence score. Typically, the value should represent a default assumption that you have no evidence that the match is a false positive.

Your organization should typically have no confidence that the match is a false positive until sufficient evidence determines otherwise. Although the goal of IMDS is to reduce the number of false positives that your organization must manually process, if your rules define a false positive too broadly, then you may miss some true matches.

Rules

A rule for IMDS contains a condition that tests a match for evidence of a false positive, a coefficient that increases or decreases the sum of evidence if the condition evaluates to TRUE for the match, and a reason code that describes the rule as part of the audit trail.

Conditi ons

A condition tests a match for evidence of a false positive by comparing an attribute against a specified value. If the attribute contains a value that satisfies the comparison, then the condition evaluates to TRUE. If the match attribute does not contain a value that satisfies the comparison, then the condition evaluates to FALSE.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 10

CONFIDENTIAL

Attribute s

An attribute represents a particular data element for an entity or a match. Rule conditions compare attributes against a specified value. Attributes can include descriptive text or numeric values such as an entity name, address, or DOB; categorical data, such as an entity gender or country; or calculated values, such as a match score or the number of days between two DOB values.

IMDS provides the following groups of attributes:

Input Attributes

Input attributes describe input entities that users search against in XG.

List Attributes

List attributes describe list entities that XG returns as matches for input entities.

Match Attributes

Match attributes describe results from the match analysis by the LexisNexis® Search Core.

Entities can be associated with multiple instances of the following data elements: names, DOB values, countries, ID numbers, addresses, citizenships, and phone numbers. The LexisNexis Search Core calculates the match score for each data element and returns a "best match" attribute and a "best score" attribute for the data element with the highest match score.

The following table lists the attributes that are available for IMDS, aligned by data element type.

Possible Attributes

Input Attributes

List Attributes

Match Attributes

List entity unique ID

Overall match confidence score Entity type conflict Best name score Sum of AKA (also known as) match scores Maximum name match score

Input entity type Full name for input entity

List entity type Best full name for list entity

DOB for input entity Calculated age for input entity

Best DOB for list entity DOB count for list entity Minimum calculated age for list entity Maximum calculated age for list entity List entity includes a DOB

Best DOB score Best DOB is partial Best DOB conflict Minimum distance in days between the best DOB and the input DOB

Gender for input entity ID number count for input entity

Gender for list entity ID number count for list entity

Gender conflict Best ID number score Best ID number conflict Sum of ID number match scores

BIS249-EN_US — 29 September 2023

IMDS in XG Page 11

CONFIDENTIAL

Possible Attributes (Continued)

Input Attributes

List Attributes

Match Attributes

Address count for input entity

Address count for list entity Best country for list entity List entity address includes a first line

Best address score Best address is partial Best address conflict Sum of address match scores Best country score Best country conflict Sum of country match scores Best country match type

Best citizenship score Citizenship conflict

These attributes are available for only LexisNexis® WorldCompliance™ Data Plus.

Phone number count for input entity

Phone number count for list entity

Best phone number score Best phone number conflict

List entity segment Search criteria codes for list entity Screening list file name for list entity Screening list reason for inclusion for list entity

Sum of additional information match scores

Coeffic ie n ts

A coefficient is the value that a rule returns when the rule's condition returns TRUE. If the condition returns FALSE, then the rule does not return a coefficient.

You use positive coefficient values for rules that test for evidence that the match is a false positive, because positive values increase the false positive confidence score. You use negative coefficient values for rules that test for evidence that the match is a true match, because negative values decrease the false positive confidence score. Larger coefficient values (farther from zero) give the rule greater weight in the false positive confidence score, and smaller values (closer to zero) give the rule less weight.

BIS249-EN_US — 29 September 2023

IMDS in XG Page 12

CONFIDENTIAL

Reason Co d es

A reason code is a brief description of the rule. If the rule is used to determine that the match is a false positive, then IMDS returns the reason code in the alert. Effective reason codes are relatively short, concise summaries of the rule condition or purpose.

Reason codes can help in the auditing process to explain why a particular match was determined to be a false positive. If your default assumption is that you have no evidence that the match is a false positive, and you test for evidence to determine otherwise, then the reason codes can provide a trail that explains what evidence for a false positive the rules found in the match.

False Pos it ive Co nfidenc e Sc or e

A false positive confidence score summarizes the likelihood that a match is an automatic false positive.

When IMDS tests a match for evidence of a false positive, each rule that evaluates to TRUE returns a numeric coefficient (or rule weight) value. IMDS adds the coefficient values to the intercept (or default assumption) value to calculate a sum that quantifies the strength of the evidence that the match is a false positive, called x.

IMDS uses the sum of evidence (x) as a negative exponent to the mathematical constant e (approximately 2.7183) in the following formula to calculate the false positive confidence score:

The following graph plots the false positive confidence score against the sum of evidence (x).

Result Curve for the False Positive Confidence Score

BIS249-EN_US — 29 September 2023

IMDS in XG Page 13

CONFIDENTIAL

The false positive confidence score never evaluates to exactly 0 or 10. Instead, as the absolute value of x increases, the value of the false positive confidence score changes by smaller amounts. For example, when x = -10, the false positive confidence score evaluates to about 0.0005. When x = -15, the false positive confidence score evaluates to about 0.000003. To simplify the result, LexisNexis Risk Solutions rounds the false positive confidence score to four decimal places. For values of x that are less than or equal to -13, the false positive confidence score rounds to 0.0000. For values of x that are greater than or equal to 13, the false positive confidence score rounds to 10.0000.

Rule E xa mp les

Rule examples can help you understand the variety of possible rules that your organization can develop.

The following rule examples provide possible use cases. The provided examples contain some rules that overlap. For example, some of the data element conflict rules are shown more than one time, with different coefficient values to show how the rules can be used in different ways. However, in a production environment, you typically only include rules that contain unique conditions. The example coefficient values work best in a rule file that contains an intercept value of -13.

Standard Automatic False Positive Example Rules

The following table contains examples of rules that test for various data element conflicts. These example rules use coefficient values of 26 to ensure that IMDS returns a false positive confidence score of 10.0000 (full confidence of a false positive) when at least one of the rules evaluates as TRUE.

Standard Automatic False Positive Rules

Reason Code

Conditions

Coefficient

DOB conflict DOB tolerance 12 months

The "best DOB conflict" flag returns TRUE. The minimum distance in days between the best DOB and the input DOB is greater than 365. The "best ID number conflict" flag returns TRUE. The "gender conflict" flag returns TRUE. The "best address conflict" flag returns TRUE. The "best country conflict" flag returns TRUE. The "best phone number conflict" flag returns TRUE. The "entity type conflict" flag returns TRUE. The "citizenship conflict" flag returns TRUE.

26 26

ID number conflict Gender conflict Address conflict Country conflict Phone number conflict Entity type conflict Citizenship conflict

26 26 26 26 26 26 26

BIS249-EN_US — 29 September 2023

IMDS in XG Page 14

CONFIDENTIAL

Automatic False Positive Override Example Rules

The following table contains examples of rules that test for an exact data element match to override other evidence of a false positive. These example rules use very large coefficient values to outweigh any possibility of a false positive decision. The precise value can be any large number, but should be at least greater than the sum of all positive coefficients in the rule file.

Automatic False Positive Override Rules

Reason Code

Conditions

Coefficient

Exact DOB match

The best DOB score is 100, AND the "best DOB is partial" flag returns FALSE. The best ID number score is 100.

-500

Exact ID number match

-500

Multiple Condition Example Rules

The following table contains examples of rules that combine multiple conditions to test for evidence of a false positive. These example rules use coefficient values of 26 to ensure that IMDS returns a false positive confidence score of 10.0000 (full confidence of a false positive) when at least one of the rules evaluates as TRUE.

Multiple Condition Rules

Reason Code

Conditions

Coefficient

Input age under 30 and former PEP (politically exposed person) with no list DOB PEP family member and country conflict

The calculated age for the input entity is less than 30, AND the "list entity includes a DOB" flag returns FALSE, AND the search criteria codes for the input entity return former PEP. The list entity search criteria codes return PEP family member, AND the "best country conflict" flag returns TRUE.

Evidence-Building Example Rules

The following table contains examples of rules that test for various data element conflicts to build up evidence for a false positive. These example rules use coefficient values of 13 to ensure that IMDS returns a false positive confidence score of 10.0000 (full confidence of a false positive) when at least two rules evaluate as TRUE If only one rule evaluates as TRUE, then IMDS returns a false positive confidence score of 5.0000 (partial confidence of a false positive).

Evidence-Building Rules

Reason Code

Conditions

Coefficient

Gender conflict Country conflict Entity type conflict Citizenship conflict

The "gender conflict" flag returns TRUE. The "best country conflict" flag returns TRUE. The "entity type conflict" flag returns TRUE. The "citizenship conflict" flag returns TRUE.

13 13 13 13

BIS249-EN_US — 29 September 2023

IMDS in XG Page 15

CONFIDENTIAL

Evidence-Building Rules (Continued)

Reason Code

Conditions

Coefficient

Name score 85-89

The best name score is greater than 84, AND the best name score is less than 90. The calculated age for the input entity is less than 30, AND the "list entity includes a DOB" flag returns FALSE, AND the search criteria codes for the list entity return former PEP.

Input age under 30 and former PEP with no list DOB

Audit Trail Example Rules

The following table contains examples of rules that do not indicate evidence of a false positive or a true match, but are recorded in the audit trail. These example rules use coefficient values of 0 to ensure that any rules that evaluate as TRUE do not affect the false positive confidence score.

Audit Trail Rules

Reason Code

Conditions

Coefficient

Name score 90-95

The best name score is greater than 89, AND the best name score is less than 96. The best name score is greater than 84, AND the best name score is less than 90. The minimum distance in days between the best DOB and the input DOB is greater than 0, AND the minimum distance in days between the best DOB and the input DOB is less than 366.

Name score 85-89

DOB is 1-365 days apart

BIS249-EN_US — 29 September 2023

IMDS in XG Page 16

About LexisNexis® Risk Solutions

LexisNexis® Risk Solutions harnesses the power of data and advanced analytics to provide insights that help businesses and governmental entities reduce risk and improve decisions to benefit people around the globe. We provide data and technology solutions for a wide range of industries including insurance, financial services, healthcare and government. Headquartered in metro Atlanta, Georgia, we have offices throughout the world and are part of RELX Group (LSE: REL/NYSE: RELX), a global provider of information and analytics for professional and business customers across industries. RELX is a FTSE 100 company and is based in London. For more information, please visit www.risk.lexisnexis.com and www.relx.com.

BIS249-EN_US — 29 September 2023